For marketers, determining the value of their marketing activities is more critical and complex than ever – especially as customers demand increased personalisation and heightened value in every brand interaction.
Using a multi-touch attribution (MTA) model can greatly assist.
The Mobile Marketing Association of SA (MMA SA) decided to drive this important conversation, by hosting international, multi-touch attribution guru, Joel Rubinson, at an event.
What is multi-touch attribution?
Multi-touch attribution is a methodology that allows marketers to really understand within their overall marketing plan that there are hundreds of different media tactics and publishers that they are working with and consumer segments that they are targeting.
Rubinson explains. “There are many combinations of that, and some of those combinations are working much better than others. And if you can find that out while the campaign is going on, you’re able to pivot during the campaign, move money from those media tactics that are not working so well to those that are working much better, and improve your overall ROI of your advertising expenditures by up to 30%.”
The future of multi-touch attribution
At present, according to Rubinson, MTA has been adopted by 40% of marketers in the US, and it’s on the increase (with the percentage expected to rise substantially), but it is still lower than some other marketing productivity methodologies.
“75% of marketers tell us that they want to use MTA … The main inhibiting factor is whether or not a marketer has their data assets put together in a way that allows for multi-touch attribution,” he comments.
He confirms that it is an expensive process to implement a MTA model, not only the analytics but also the data-sets. Despite being a powerful methodology, MTA works off of user-level data, which requires a certain level of sophistication. Connecting ad serving information and conversation data at a user level is required. This can be complicated and costly, but the returns are well worth the time and monetary investment.
Rubinson affirms that it is hard work and costly, but then motivates for it by citing a case study where a MTA model was used to determine that one segment delivered 16 times the return on ad spend than another segment. MTA allows marketers to truly know who consumers are and what segments they fall in. A very rich profile, if a marketer has the right data assets, can be created for a segment or a persona.
“Over time, marketers will rise to that challenge and adopt the methodology, because 20 to 30% profit growth by using the right method is not something that you can pass up, you need to engage with that. How are you going to throw away growth points? That’s what you get judged on,” says Rubinson.
Using the right partners in the process and being smart and innovative about data gathering techniques are some of the ways to negate the expenses.
While MTA humanises and personalises the marketing process, it does so in a way that doesn’t flout privacy laws. “The user-level data for multi-touch attribution is always anonymised. By the time it gets to the analyst who’s running the models it’s always anonymised, without possibility of reidentification. Those are the rules of the game,” Rubinson reassures.
History of multi-touch attribution
MTA is not new, having been around for four or five years, with some degree of prominence. Attribution existed before that in a broader sense and there are still simpler forms today (examples include last click attribution and first touch attribution).
Rubinson describes MTA as “a more scientific, statistically robust method that’s been proven to be much more accurate at assigning credit properly”, adding that “MTA is able to get underneath channel-level investments, exploring relationships to sales”.
In earlier generations of attribution models, the consumers were not in the equation. They explored the amount being spent and the sales figures, with no inclusion of consumers. In MTA, it’s all about consumers, as it’s user-level consumer data.
Here are some pictures from the event:
Follow Michael Bratt on Twitter @MichaelBratt8